Objectives
This work package (WP2000) represents a critical transition from technical specification to implementation. It focuses on identifying, refining, and evaluating the most promising AI-based forecasting solutions for GOVSATCOM traffic prediction. WP2000 aims to establish a technical baseline, quantify trade-offs, and validate the feasibility and compliance of each solution, preparing the groundwork for demonstrator deployment in later phases.
-
The specific technical objectives of WP2000 -- Proof-of-concept for the selected technical solutions are:
-
Objective 1: Identify and define viable forecasting models. Based on the outputs of WP1000, CTTC selects several AI-based forecasting solutions and backup options for GOVSATCOM traffic prediction, considering their theoretical and operational potential.
-
Objective 2: Perform trade-off analysis across technical solutions. In collaboration with EUTELSAT, a detailed comparison is conducted between the selected techniques. Metrics include accuracy, robustness, computational cost, explainability, and adaptability to satellite operational requirements.
-
Objective 3: Perform trade-off analysis across technical solutions. In collaboration with EUTELSAT, a detailed comparison is conducted between the selected techniques. Metrics include accuracy, robustness, computational cost, explainability, and adaptability to satellite operational requirements.
-
Objective 4: Justify final solution(s) and evaluate feasibility. Each model is validated against key operational parameters and constraints. The feasibility assessment includes scenario-based testing and critical parameter analysis.
-
Objective 5: Quantify expected performance and assess compliance. Models are evaluated in terms of their forecast accuracy, responsiveness, and alignment with GOVSATCOM compliance requirements, generating a provisional statement of compliance for each PoC.
